Overview: The article discusses the significance of inertial measurement units in human motion analysis, highlighting their role in tracking movement. It explains the key components and addresses challenges in various applications.

An important tool for evaluating movement parameters is human motion analysis. This is particularly important for assessing exercise routines, medical conditions, and health promotion. It plays a vital role among elderly people who want to track their physical activities. There will be a greater need for home-based rehabilitation programs and independent access to exercise data by each individual to track their medical condition.

Optical methods accurately assess both kinematic data and motion data, and they are regarded as the gold standard in the field of motion analysis. However, they have several drawbacks, including expensive equipment, the need for skilled individuals to operate them, extensive installation spaces, and reduced moveability that restricts their application only indoors.

Inertial motion analysis has become a viable substitute for optical approaches. An advantage of inertial systems over optical systems is that they are compact, reasonably priced, and portable. Because of this, inertial measurement units (IMUs) are a user-friendly option for estimating human movements. 

These devices can continuously monitor human movements during each activity, and they are more accurate than the data collected from occasional laboratory testing. For these reasons, IMU use for continuous human motion tracking has increased during the past few decades.

What is an inertial measurement unit?

IMU is a microelectromechanical system (MEMS) sensor that measures the object’s specific force, angular rate, and orientation. These devices monitor human lives during their various routines and record changes in location, body movements, and rotational changes.

Fig. 1: MPU-9250 by InvenSense (Accelerometer, Gyroscope, Magnetometer) Source: oemsecrets

Key Components of an Inertial Measurement Unit

It combines various sensors, including an accelerometer, gyroscope, and magnetometer. Three axial magnetometers, three axial gyroscopes, and three axial accelerometers make up the typical inertial measurement unit, as shown in Fig. 2.

Fig. 2: Inertial measurement unit Source: Sciendo

Accelerometer

Linear acceleration resulting from the applied force can be measured using an accelerometer.

Gyroscope

Angular velocity and rotational movement can be tracked using a gyroscope.

Magnetometer

The orientation of the sensors based on the Earth’s magnetic field can be measured with a magnetometer. It supports the other two sensors in determining the direction and magnitude of motion.

The use of accelerometers alone is more common (8.8%). Their particular force measurement enables them to directly observe the gravity vector, which is used as a point of orientation. However, the gravity vector can be directly observed in the normal gait cycle phases only when accelerometers are static.

Magnetometers are the most limited sensors examined due to their extreme sensitivity to magnetic changes in the environment. Therefore, the use of this sensor independently is limited.

Techniques for fusing sensors are helpful to overcome the challenges faced by each sensor individually. These IMU sensor data can be used to determine a device's 3-D position. The gyroscope, magnetometer, and accelerometer work together to provide information on the angular rate of motion as well as vector gravity and Earth's magnetic field references. Sensor fusion techniques using various combinations of the three sensors can also predict these 3-D positions. 

Working of the Inertial Measurement Unit

IMUs detect linear acceleration and rotational rate through accelerometers and gyroscopes. The data collected from these sensors is often integrated using algorithms to provide comprehensive motion and orientation information. In some configurations, a magnetometer is used for heading reference, enhancing the accuracy of orientation data.

Application

IMUs may be manufactured at a comparatively low cost but with higher accuracy, significantly expanding the IMU's application in many new fields like indoor localization, sports analysis, and body activity classification. IMUs are employed in many different fields, including healthcare, sports, and navigation of diverse platforms, including autonomous underwater vehicles, unmanned land vehicles, aerial vehicles, and smartphones. 

These IMU sensors are of great use in the sports industry in tracking exercise, swimming, posture, etc. The healthcare industry employs them increasingly to monitor physical and physiological activities. For instance, sudden changes in posture and position can be tracked with these sensors. Prompt action can be taken in the case of elderly patients, and activities like fall incidents can be avoided. 

Challenges

IMU measurements have challenges regarding the accuracy of the data collected, which results in drift over time. They are merged with external sensors to avoid drift. Using a multiple IMU (MIMU) design is one way to reduce drift while keeping the system affordable. They offer improved dynamic measurement range, accuracy, and dependability compared to a single IMU. Additionally, using accelerometer data, the MIMU architecture can directly estimate angular acceleration and motion. 

Secondly, with significantly lower costs and higher precision, inertial sensors can be produced as MEMS systems. Systematic and stochastic errors can affect MEMS inertial sensors. 

Inaccurately calibrated systems generate systematic errors, which can be minimized by requiring extremely careful calibration. Practically, a series of parameters causes systematic errors, such as:

  • Nonzero biases
  • Cross-axis sensitivities
  • Erroneous scaling
  • Sensor axis misalignment

The initial data from a sensor with systematic errors can be changed into more precise information without systematic errors with these parameters and matrix multiplication procedures. Many optimization algorithms, including the least squares method and the Kalman filter, are frequently used to optimize these parameters.

Thus, IMU calibration is one of the most important steps in lowering systematic errors. On the other hand, stochastic errors cannot be avoided because they result from measurement errors. 

To conclude, IMU is vital in many modern technologies, providing essential data for navigation, orientation, and motion tracking. Despite their limitations, their integration with other technologies continues to expand their applicability, making them indispensable in fields ranging from consumer electronics to aerospace and robotics.

Summarizing the Key Points

  • Inertial Measurement Units are crucial for tracking human motion, offering applications in healthcare, sports, and rehabilitation, and enhancing the accuracy of movement analysis and monitoring.
  • IMUs consist of accelerometers, gyroscopes, and magnetometers, which provide comprehensive data on linear acceleration, angular velocity, and orientation in various environments.
  • Integrating multiple IMUs (MIMUs) can improve accuracy and reduce drift, enhancing dynamic measurement range and dependability in applications like indoor localization and body activity classification.

Reference

Sara García-De-Villa et al., “Inertial Sensors for Human Motion Analysis: A Comprehensive Review,” IEEE Transactions on Instrumentation and Measurement 72 (January 1, 2023): 1–39, https://doi.org/10.1109/tim.2023.3276528

Ariel Larey, Eliel Aknin, and Itzik Klein, “Multiple Inertial Measurement Units–An Empirical Study,” IEEE Access 8 (January 1, 2020): 75656–65,
https://doi.org/10.1109/access.2020.2988601

Xin Zhang et al., “Low-Cost Inertial Measurement Unit Calibration With Nonlinear Scale Factors,” IEEE Transactions on Industrial Informatics 18, no. 2 (February 1, 2022): 1028–38, https://doi.org/10.1109/tii.2021.3077296

Dariusz Tomaszewski, Jacek Rapiński, and Renata Pelc-Mieczkowska, “Concept of AHRS Algorithm Designed for Platform Independent Imu Attitude Alignment,” Reports on Geodesy and Geoinformatics 104, no. 1 (December 20, 2017): 33–47,
 https://doi.org/10.1515/rgg-2017-0013